strisunshinewentingwang/pytorch

Pull Requests Commits

[c10d][Fix] Start gloo sequence numbers at 0. (#101422)

rohan-varma committed 1 year ago

6f8a71aa

[IValue] Only coalesce once (#101447)

rohan-varma committed 1 year ago

4b849744

Revert "[fx] Better replacements finder in subgraph rewriter (#100556)"

pytorchmergebot committed 1 year ago

13056ca2

Add more canonical way of adding runtime pass (#100956)

tugsbayasgalan committed 1 year ago

194d3603

Use %zu instead of %ld when formatting size_t (#101412)

pandaninjas committed 1 year ago

f0786ad7

Clean up grad check in sdp_utils.h (#101435)

drisspg committed 1 year ago

52363de2

Use python3 instead of /usr/bin/env python3 on Windows (#101437)

huydhn committed 1 year ago

c3f7db3f

Fix Wishart distribution documentation (#95816)

alexisthual committed 1 year ago

24cc7fe0

[Inductor] Improve view/reshape on tensors with shape 0 (#101051)

yanboliang committed 1 year ago

7f3b00bf

Revert torch.fx.interpreter error printing change (#101462)

ezyang committed 1 year ago

d1980336

[caffe2/tools/autograd] Fix non-determinism in code gen (#101425)

andrewjcg committed 1 year ago

799ef7e5

fix print tensor in cpp for privateuse1 (#100797)

heidongxianhua committed 1 year ago

a8376099

[decomp] fix decomp of batch_norm when weight/bias is not flattened (#101059)

jgong5 committed 1 year ago

788ff062

Inductor cpp wrapper: support sympy.Expr as input (#101257)

chunyuan-w committed 1 year ago

1faef895

Enable default workflow PyT 2.0 UTs on ROCm stack (#100981)

jataylo committed 1 year ago

187eb7ca

[opinfo] empty_strided (#100890)

khushi-411 committed 1 year ago

01c71065

Apply the same fix to cleanup process on Windows CPU build job (#101460)

huydhn committed 1 year ago

3920ec14

Rename minpybind namespace from py to mpy (#101410)

ezyang committed 1 year ago

0577043d

[small BE] update NcclTest dim size (#101127)

H-Huang committed 1 year ago

a206e8b0

Update cpp_extension.py (#101285)

pminimd committed 1 year ago

59a3759d

Bump up flatbuffer submodule version to the latest release (v23.3.3) (#100716)

zasdfgbnm committed 1 year ago

17320777

[PTD] Make input contiguous for _ReduceScatter (#101373)

fduwjj committed 1 year ago

9d858642

[dynamo 3.11] enable other torch 3.11 dynamo-related tests (#99180)

williamwen42 committed 1 year ago

0e811044

Disable locality reodering in training (#101423)

eellison committed 1 year ago

3b7c6b21

Fix checkpoint doc small formatting issue (#101419)

soulitzer committed 1 year ago

70ef0bb4

[SPMD] Allow Override.replacement to have a global view (#101427)

mrshenli committed 1 year ago

af841f38

Remove input tracker from runtime assertion pass (#100955)

tugsbayasgalan committed 1 year ago

9ffad5b6

Fix inconsistent torch.nn.MaxPool1d output on cpu and gpu (#99843)

tfsingh committed 1 year ago

563d8058

[pt2] add meta and `SymInt` support for `linalg_lu` (#101372)

nkaretnikov committed 1 year ago

9eb1748b

[pt2] add meta for `linalg_ldl_solve` (#101367)

nkaretnikov committed 1 year ago

ac4cc63a

Newer Older